the effect of water-induced stress to enhance hydrocarbon ... · enhance hydrocarbon recovery in...
TRANSCRIPT
-
The Effect of Water-Induced Stress ToEnhance Hydrocarbon Recovery
in Shale ReservoirsPerapon Fakcharoenphol, Sarinya Charoenwongsa, Hossein Kazemi, and Yu-Shu Wu, Colorado School of Mines
Summary
Waterflooding has been an effective improved-oil-recovery (IOR)process for several decades. However, stress induced by water-flooding has not been well studied or documented. Water injectiontypically increases reservoir pressure and decreases reservoirtemperature. The increase in reservoir pressure and decrease inreservoir temperature synergistically reduce the effective stress.Because of such decrease in stress, existing healed natural frac-tures can be reactivated and/or new fractures can be created. Simi-lar effects can enhance hydrocarbon recovery in shale reservoirs.
In this paper, we calculated the magnitude of water-injection-induced stress with a coupled flow/geomechanics model. To eval-uate the effect of water injection in the Bakken, a numerical-simu-lation study for a sector model was carried out. Stress changescaused by the volume created by the hydraulic fracture, waterinjection, and oil production were calculated. The Hoek-Brownfailure criterion was used to compute rock-failure potential.
Our numerical results for a waterflooding example show thatduring water injection, the synergistic effects of reservoir coolingand pore-pressure increase significantly promote rock failure,potentially reactivating healed natural macrofractures and/or cre-ating new macrofractures, especially near an injector. The rockcooling can create small microfractures on the surface of the ma-trix blocks. In shale oil reservoirs, the numerical experimentsindicate that stress changes during water injection can improve oilrecovery by opening some of the old macrofractures and creatingnew small microfractures on the surface of the matrix blocks topromote shallow water invasion into the rock matrix. Furthermore,the new microfractures provide additional interface area betweenmacrofractures and matrix to improve oil drainage when using sur-factant and CO2 enhanced-oil-recovery techniques. These positiveeffects are particularly important farther away from the immediatevicinity of the hydraulic fracture, which is where much of theundrained oil resides.
Introduction
Waterflooding has been an effective IOR process for decades.However, stress induced by changes in pressure or temperatureduring waterflooding is typically overlooked or not adequatelyreported. Water injection changes the stress field in the reservoirand the surrounding rocks by changes in pore pressure and reser-voir temperature. For instance, the change in stress within a reser-voir can be estimated with the following equation for isotropicrock.
rh � ap ¼t
1� t ðrv � apÞ þE
1� t2 ðeh þ teHÞ
þ E1� t ½bðT � T0Þ� � � � � � � � � � � � � � � � � ð1Þ
In Eq. 1, rh and rv are the minimum horizontal and verticalstresses, respectively; a is Biot’s coefficient; b is the linear thermal
expansion; � is Poisson’s ratio; p is pore pressure; T is tempera-ture; T0 is initial temperature; and eh and eH are the strain in theminimal- and maximal-horizontal-stress directions, respectively.
The first term on the right side of Eq. 1 is the contributionfrom the vertical stress to the effective-horizontal-stress change.The second term is the contribution of the horizontal mechanicalstrain, and the last term is the contribution from the thermal strain.By use of the classical uniaxial-strain assumption, we drop thesecond term of Eq. 1, which yields the following equation:
rh � ap ¼t
1� t ðrv � apÞ þE
1� t ½bðT � T0Þ�: ð2Þ
With Eq. 2, we can derive the following equations, which rep-resent the change in effective stresses as a function of the changein pore pressure and reservoir temperature:
ðr0h � r0h0Þ ¼ �t
1� t aðp� p0Þ þE
1� t bðT � T0Þ ð3Þ
ðr0v � r0v0Þ ¼ �aðp� p0Þ; ð4Þ
where p0 is the initial reservoir pressure, r0h ¼ rh � ap, andr0h;0 ¼ rh;0 � ap0.
As for an application of Eq. 3, consider the following example.Suppose we inject cold water into a reservoir such that theincrease in pore pressure is 1,000 psia, whereas the decrease intemperature within the drainage volume of the well is 50�F. Thedata pertinent to this problem are the following: a¼ 0.8, �¼0.2, E ¼ 2:83� 106 psi (18 GPa), and b ¼ 5:56� 10�6 �F�1(10�5K�1).
The change in effective horizontal stress because of pore-pres-sure increase of 1,000 psia is �200 psia, and the effect of coolingby 50�F is �912 psia, for a net effective-horizontal-stress changeof �1,112 psia. The effective-vertical-stress change is �800 psi.The Mohr diagram, Fig. 1, shows that the synergistic effects ofreservoir cooling and pore-pressure increase promote rock failure,potentially reactivating healed natural fractures and creating newmicrofractures.
A recent waterflood pilot test in the Viewfield Bakken by Cres-cent Point Energy (Wood and Milne 2011) indicates improved oilproduction. The first pilot test included four horizontal producersand one horizontal injector, drilled parallel to each other withwell spacing of approximately 656 ft (200 m) (Fig. 2a). The firstpilot saw a robust production response through the latter part of2008 and into much of 2009, as shown in Fig. 2b. After reachinga peak production rate of approximately 550 B/D, productiondeclined by approximately 25% over the next 2 years. Thisdecline rate is a great improvement over the 70-to-75% ratedecline in the first year of a typical declining rate in the ViewfieldBakken wells (Wells 2011).
Crescent Point Energy also conducted a tracer test. The resultsindicated immediate water breakthrough to the offset producers.This is not a surprise because injected water tends to flow throughthe fracture networks created by hydraulic fracturing, displacingthe oil toward the production well, and bypassing tight shale ma-trix, rapidly reaching producers.
We believe that water-injection-induced stress enhances per-meability of existing healed natural fractures and creates new
. . . . . . .
. . . .
. . . . . . . . . . . . . . . . . . . . .
Copyright VC 2013 Society of Petroleum Engineers
This paper (SPE 158053) was accepted for presentation at the SPE Annual TechnicalConference and Exhibition, San Antonio, Texas, USA, 8–10 October 2012, and revised forpublication. Original manuscript received for review 18 June 2012. Revised manuscriptreceived for review 26 November 2012. Paper peer approved 16 January 2013.
October 2013 SPE Journal 897
-
microfractures penetrating into tight matrix rock. These micro-fractures create flow paths for hydrocarbons inside the matrix,thus improving the fracture/matrix interface area and increasinghydrocarbon production from the matrix. In addition, the princi-pal mechanism of oil displacement by waterflooding is mainlyby viscous displacement of oil from the microfractures andpossibly with a little contribution from the matrix. The theoreti-cal support for this conclusion will be provided in a futurepublication.
Because waterflooding appears to be a viable IOR technique,in this paper, we further examine the role of waterflood-inducedstress to improve oil recovery. We calculate the magnitude of
water-injection-induced stress with a coupled flow/geomechanicsmodel.
In addition to the preceding engineering analysis, we present anumerical model to describe how waterflooding improves oil re-covery in shale reservoirs. In support of our numerical-modelingstudy, published experimental data relating to the waterflood-induced stress will be presented and discussed.
Thermal Stimulation in Geothermal Reservoirs
Thermal stimulation has been reported as a successful stimulationtechnique for geothermal reservoirs (Siratovich et al. 2011). Wespeculate that this mechanism is also a key to success in a cold-waterinjection in Bakken. A typical thermal stimulation is conducted byinjecting cold water into hot reservoir rock. Three mechanisms areconsidered for the success: the reopening of old fractures by thermalcontraction, the creation of new fractures from temperature-inducedstress exerted on reservoir rocks, and the cleaning of debris or min-eral deposits from open fractures (Axelsson et al. 2006).
Siratovich et al. (2011) thoroughly review thermal-stimulationapplication in a number of geothermal fields worldwide. The stim-ulation could be conducted right after a well is drilled or long af-ter production, and often is cyclic. The treatment could beperformed in a short period ranging from hours to days. Thermalcycling with rapid cooling and heating of the wellbore (by inject-ing cold water and stopping injection) shows positive resultsenhancing near-wellbore permeability (Kitao et al. 1990; Bjorns-son 2004). Siratovich et al. (2011) also conducted experiments to
(a)
(b)
600
500
400
300
200
Pro
d’n
(bbl
/d)
Inj’n
(m
bbl/d
)
Daily Prod’n
Feb
-06
Feb
-08
Feb
-09
Feb
-10
Feb
-07
Jun-
06
Jun-
08
Jun-
09
Jun-
10
Oct
-09
Oct
-10
Jun-
07
Oct
-06
Oct
-08
Oct
-07
Cummn Prod’n
Cum
mn
Pro
d’n
(mbb
l)
100
0
600
Pilot 1Well ID Lat. Length
1 191/13-10-008-09W2/00 1500 m1385 m1391 m1510 m1385 m5 191/04-10-008-09W2/00
4 191/05-10-008-09W2/003 191/12-10-008-09W2/002 191/12-10-008-09W2/00
–1605.85
0.00(–m)
ProducerInjector
405 m200 m205 m
400 m1
23
45
Injector E
S
N
W
500
400
300
200
100
0
4
20
Fig. 2—Waterflooding Pilot Test No. 1 in the Viewfield Bakken by Crescent Point Energy Corporation: (a) well configuration and (b)oil-production and water-injection profiles (Wood and Milne 2011).
Shear
stre
ss
Increase pressureInta
ct rock
Healed
fracture
d rock
Decrease temperature
Effective normal stress, σ – ap
σ3,2 σ3,1 σ1,2 σ1,1' ' ' '
Fig. 1—Mohr diagram for stress change during water injection.
898 October 2013 SPE Journal
-
verify that temperature-induced stress can create fractures in avolcanic-rock specimen. The experiment was carried out by heat-ing the specimen to 300 to 650�F and immediately submergingthe specimen in a 68�F water bath. Newly created fracturesobserved on the specimen indicate that temperature-induced stresscould create new microfractures (Fig. 3a).
Moreover, numerical studies (Ghassemi and Zhang 2004;Ghassemi et al. 2007; Tarasovs and Ghassemi 2012) also illustratethe role of temperature-induced stress to create new fractures.Cold-water circulation in hot fractured rock induces thermal con-traction and creates tension near main fractures. Once the inducedstress exceeds the rock strength, secondary fractures can propa-gate from the main fractures (likely perpendicular to them) intomatrix rock (Fig. 3b).
We speculate that water-induced microfractures could occurduring water injection in the Bakken. These induced microfrac-tures can enhance oil recovery by increasing macrofractures/matrix interface area, thus improving oil production from tightBakken reservoirs. However, the Bakken reservoir temperature isnot as high as that of the aforementioned geothermal reservoirs; itwould take a longer period to see the effect of the thermal stimu-lation. This is consistent with the observations from the water-injection pilot test by Crescent Point Energy. The company indi-cated that the improvement from water injection could be detectedseveral months to 1 year after starting injection (Wells 2011).
The Research Objective
The objective of this study is to investigate the role of pressureincrease and temperature decrease during water injection in thestress redistribution in conventional and unconventional reservoirsand the tendency of water-induced microfractures to form in theBakken. A coupled flow/geomechanics model is used to calculatepore-pressure- and temperature-induced stress in the reservoirs. Inthis paper, we conduct two simulation studies: five-spot patternwaterflooding in a conventional reservoir and water injection in amultistage hydraulic-fractured horizontal well in the Bakken.
Mathematical Model for Coupled Flow andGeomechanics
A coupled flow and geomechanics model is used to describe fluidflow, heat transfer, and rock deformation in the formations. Thismodel is built on the basis of the governing equations preservingphysical laws (including conservation of mass, momentum, andenergy) and constitutive laws (such as Darcy’s law, porothermo-elasticity, and infinitesimal strain theories). In this paper, weextend the work by Charoenwongsa et al. (2010) to capture a me-chanical anisotropic system. Detailed implementation can be foundin Charoenwongsa et al. (2010).
Elements of the model include the pore-pressure equation,water-saturation equation, temperature equation, rock-displace-ment equation, and stress/strain relation.
Pore Pressure Equation.
�r �~vw �r �~v/ þ q̂t
¼ ð1� evÞ/ ct@p/@t� bfl;t
@T
@t
� �� / @ðr �~uÞ
@t; � � � � � � ð5Þ
where ~v is Darcy velocity; q̂ is sink and source term per unit vol-ume; subscripts o and w represent the oil and water phases,respectively; ev is volume matrix strain; / is porosity; ct is totalcompressibility; bfl;t is total thermal expansion of fluid; t is time,~u is the rock displacement vector defined as ~u ¼ ½u; v;w�T , andu; v; and w are rock displacement components in the x-, y-, andz-directions, respectively.
Water Saturation Equation.
�r �~vw þ q̂w
¼ ð1� evÞ/ Swðcw þ c/Þ@p
@t� Swbfl;w
@T
@tþ @Sw
@t
� �
� /Sw@ðr �~uÞ
@t: � � � � � � � � � � � � � � � � � � � � � � � � � ð6Þ
where c/ and cw are pore and water compressibility, respectively.
Temperature Equation.
� r � ðcp;wqw~vw þ cp;oqo~voÞðT � TnÞ� �
þ r � ðkTrTÞþ ðcp;wqwqwÞðTinj � TnÞ� �
¼/ðcv;wqwSw þ cv;oqoSwÞþð1� /Þcsqs
� �@T
@t
�3bKbT0@ðr �~uÞ
@t
;
� � � � � � � � � � � � � � � � � � � � � � � ð7Þ
where q is density; cp is constant pressure heat capacity; cv is con-stant-volume heat capacity; subscripts o and w represent the oiland water phases, respectively; cs is heat capacity of solid, and Kbis drained bulk modulus.
Rock Displacement Equation. X-Direction.
@
@xC11
@u
@xþ C12
@v
@yþ C13
@w
@z
� �
þ @@y
C44@u
@yþ @v@x
� �� �þ @@z
C66@u
@zþ @w@x
� �� �
¼ �a @po@x� fw
@pcwo@x� ð fwcw þ focoÞ
@D
@x
� �
�bðC11 þ C12 þ C13Þ@T
@x� � � � � � � � � � � � � � � � � � � ð8Þ
(a) (b)
Injection well
Secondary thermal cracks
Extraction well
Fig. 3—Recent studies confirming thermal induced fractures in geothermal reservoirs: (a) an experimental study by Siratovichet al. (2011) showing fracture creation during the experiment on Tjaldfell Tholeiitic basalt specimen by heating the sample to 400ºFand immediately submerging it in a 70ºF water bath and (b) a numerical study by Ghassemi (2012) showing secondary fracturescreated by cold-water injection in geothermal reservoirs.
October 2013 SPE Journal 899
-
where Cij represents rock mechanical properties defined in astrain/strain relation under drained conditions; detailed calculationis given in Appendix A. pcwo represents capillary pressure of awater/oil system, f is fractional flow, c is the pressure gradient,and D is depth.
Y-Direction.
@
@xC44
@v
@xþ @u@y
� �� �þ @@y
C21@u
@xþ C22
@v
@yþ C23
@w
@z
� �
þ @@z
C55@v
@zþ @w@y
� �� �
¼ �a @po@y� fw
@pcwo@y� ð fwcw þ focoÞ
@D
@y
� �
�bðC21 þ C22 þ C23Þ@T
@y: � � � � � � � � � � � � � � � � � � � ð9Þ
Z-Direction.
@
@xC66
@w
@xþ @u@z
� �� �þ @@y
C55@w
@yþ @v@z
� �� �
þ @@z
C31@u
@xþ C32
@v
@yþ C33
@w
@z
� �
¼ �a @po@z� fw
@pcwo@z� ð fwcw þ focoÞ
@D
@z
� �
�bðC31 þ C32 þ C33Þ@T
@z: � � � � � � � � � � � � � � � � � � � ð10Þ
The stress/strain relation is
Drx � aðp� p0ÞDry � aðp� p0ÞDrz � aðp� p0Þ
DsxyDsyzDszx
2666666664
3777777775¼
C11 C12 C13
C21 C22 C23
C31 C32 C33
C44
C55
C66
2666666664
3777777775
�
ex þ bðT � T0Þey þ bðT � T0Þez þ bðT � T0Þ
exyeyzezx
2666666664
3777777775; � � � � � � � ð11Þ
where Dri is normal stress change on the i-direction, Dsij is shearstress change in the i, j plane, ei is normal strain on the i-direction,and eij is shear strain on the i, j plane.
For the strain definition, we use the small strain definition,which is mathematically expressed by
ex ¼@u
@xey ¼
@v
@yez ¼
@w
@z
exy ¼1
2
@u
@yþ @v@x
� �eyz ¼
1
2
@v
@zþ @w@y
� �ezx ¼
1
2
@w
@xþ @u@z
� �:
� � � � � � � � � � � � � � � � � � � ð12Þ
Water-Injection-Induced Stress
Cold-water injection reduces effective stress in reservoirs in twoways, which is evident by reviewing Eq. 1. This equation clearlyindicates that the effective stress is reduced when pore pressureincreases. Second, the effective stress is reduced when the reser-voir cools, which causes rock matrix to contract (volumetricstrain). The latter effect is similar to cooling a steel rod with twoends fixed where the volume change caused by thermal contrac-tion creates internal strain in the rod, thus decreasing the stress aswell as the effective stress.
To study water-injection-induced stress, we modeled water-flooding in a 160-acre, 50-ft-thick five-spot pattern. Thus, weshow the results for only a quarter of this five-spot reservoir seg-ment with injector-to-producer spacing of 1,867 ft. Cold water isinjected into the reservoir for 1 year. A coupled fluid-and-heat-flow/geomechanics model based on the work of Charoenwongsaet al. (2010) was used to calculate stress change in the reservoir.Fig. 4 shows the model configuration with overburden, underbur-den, and sideburden boundaries. Detailed reservoir data are pre-sented in Appendix B.
Table 1 shows the simulation results for pressure, temperature,and effective-stress changes at various locations. Here, stresschange is calculated as the change from the initial condition. Nega-tive stress change indicates reduction of the effective stress. Wecan see that temperature reduction after 1 year of injection is con-fined to a range up to 300 ft from the injection well. Thus, the tem-perature change is relatively slow compared with the pore-pressurechange. The contribution of each term on the right side of Eq. 1 tothe effective-horizontal-stress change is shown in Table 1. We canclearly see that the stress change at the vicinity of the injector isdominated by thermal strain because of the temperature decrease.The thermal-strain effect reduces farther away from the injectorbecause the temperature reduction occurs only near the injector.
The stress profiles are plotted as the Mohr circle, shown in Fig.5a, including the Mohr-Coulomb failure envelope with a 100-psicohesion and a friction angle of 30� for healed natural fractures.
Study area
Producer
Injector
Y-d
irect
ion,
ft
0
1000
2000
3000
4000
5000
X-direction,ft010002000300040005000
Sideburden
Reservoir
Dep
th, f
t
x-direction, ft y-dire
ction
, ft
02000
40000
2000
4000
0
2000
4000
6000
8000
40 acre Overburden
Underburden
Reservoir
(a) (b) (c)
Fig. 4—Model configuration: (a) study area of a quarter-model of a five-spot waterflood pattern; (b) top view of the model reservoirand sideburden; and (c) overburden and underburden layers.
900 October 2013 SPE Journal
-
On this plot, the intersection of a stress profile and the failure enve-lope indicates a possible reactivation of healed natural fractures.We can see that the maximum stress reduction is near the injector,where the maximum temperature reduction takes place. The stressand temperature changes become less pronounced farther awayfrom the injector.
Fig. 5b shows a comparison between pore pressure-inducedand temperature-induced stress changes at 200 ft from the injec-tor. The pore-pressure increase raises the total stress and causes aslight mechanical strain in the system. This increase reduceseffective stress and shifts the Mohr circle to the left. The inducedstrain creates shear stress and slightly enlarges the Mohr circle.The temperature decrease causes negative strain or tension in thesystem, and creates significant shear stress caused by the mechan-ical strain contrast in vertical and horizontal directions. Ulti-mately, this reduction widens notably and shifts the Mohr circleto the left. Both the pore-pressure increase and the temperaturedecrease move the Mohr circle toward the failure envelope. Inother words, both effects synergistically promote rock failure.
Numerical Simulation of Water Injection in aSector Model of the Bakken
To evaluate the effect of water injection in a typical Bakken reser-voir, a numerical-simulation study for a sector model was carriedout. The model represents a reservoir section between a horizontalwater injector and an oil producer in the middle Bakken formation.Both wells are assumed to be stimulated by a set of multistage hy-draulic fractures, with well spacing of 650 ft, fracture half-lengthof 150 ft, and fracture spacing of 400 ft (Fig. 6a). A typical aver-age hydraulic-fracture width, in shale formation, of 0.2 in. isassumed. The model is initialized when the hydraulic-fracturingjob has been completed. Thus, the fracture network created by thehydraulic-fracturing process exists at the start of water injection.
For simplicity, we assume that the fractures network, created asa result of hydraulic fracturing, is parallel with the principal-stresscoordinate axes, as shown in Fig. 6c. We realize that the mostaccurate way to simulate the fractures network is to model the frac-tures at an angle approximately 20 to 30� from the principal-stressaxes. Our approach simplifies computation but it does not detractfrom the main objective of the research. Fundamentally, we aremore interested to know to what extent the pore-pressure and tem-perature change in the fractures/matrix system affects the stressredistribution in the reservoir. Furthermore, our representation ofthe microfractures network is amenable to logarithmic-grid distri-bution, which leads to accurate numerical results (Fig. 6d). Macro-fractures were represented by a set of high-permeability orthogonalgrid cells with conductivity of 1.0 md-ft in conductivity, comparedwith 0.002 md-ft for the adjacent matrix grid array.
In this paper, we did not include the mineralization processduring water injection. But we suspect that cooling of the forma-tion brine should lead to some mineral precipitation, which can beincluded in the future research effort.
Overburden and underburden layers are included to capturestress redistribution because of mechanical strain (Fig. 6b). Theeffect of fracture-face displacement is modeled by applying rockdisplacement at the boundary of the model, where the hydraulicfractures are. The other lateral and bottom boundaries are assumedzero-displacement, and the surface boundary is modeled by con-stant-stress boundary (Fig. 6c).
Reservoir and rock-mechanics properties and the initial reser-voir conditions are taken from the literature and are summarizedin Appendix B (Nottenburg et al. 1978; Ye 2010; Kurtoglu et al.2011; Wang and Zeng 2011). The Bakken reservoir parametersfrom various authors reported in Appendix B are widely different.Nonetheless, because the main rock-mechanics properties aretaken from Wang and Zeng (2011), we used the information fromthe latter for consistency.
0 500 1000 1500 2000 2500 3000 3500 4000 4500 50000
500
1000
1500
2000
2500
3000
Normal stress, psi
She
ar s
tres
s, p
si
0 500 1000 1500 2000 2500 3000 3500 4000 4500 50000
500
1000
1500
2000
2500
3000
Normal stress, psi
She
ar s
tres
s, p
si
(a) (b)
At the injector
200 ft away from the injector
300 ft away from the injector
Initial condition
Temperature-induced stress
Initial condition Total induced
stress Pore pressure-induced stress
Fig. 5—Simulation results of stress changes after 1 year of injection: (a) effective-stress profile at several locations and (b) com-parison between temperature- and pore-pressure-induced stress at 200 ft away from the injector.
TABLE 1—SIMULATION RESULTS OF INDUCED PRESSURE, TEMPERATURE, AND STRESS
Change
At the
Injector
200 ft Away
From the
Injector
300 ft
Away From
the Injector
Pressure changes, p–p0 (psi) 895.7 734.9 666.3
Temperature change, T–T0 (�F) �114.7 �31.7 �7.3
Effective-stress change in vertical direction, r0v � r0v0 (psi) �1,395 �626 �456Mechanical strain in x-direction 0.000172 �3.6�10�5 �5.2�10�5
Mechanical strain in y-direction 0.000172 �3.6�10�5 �5.2�10�5
Effective-stress change in horizontal direction, r0h � r0h0 (psi) �4,365 �1,621 �653Effect of the effective-vertical-stress change on the horizontal-stress change (psi) �465 �209 �152Effect of the horizontal mechanical strain on the horizontal-stress change (psi) 689 �144 �208Effect of the thermal strain on the horizontal-stress change (psi) �4,588 �1,269 �293
October 2013 SPE Journal 901
-
The middle Bakken is modeled by an isotropic material model,but the upper and lower Bakken are transversely anisotropic mate-rial. Ye (2010) shows that the mechanical properties of the middleBakken can be represented as an isotropic material, but the upperand lower Bakken are transversely anisotropic in the principaldirections parallel and perpendicular to the bedding. Vernik andLandis (1996) describe the Bakken formation as a sequence ofthin layers where silty kerogen-depleted shale layers alternatewith kerogen-rich coal-like layers.
Rock Failure
Rock failure occurs when shear stress circle crosses the Mohr-Coulomb or Hoek-Brown failure envelope. Pore-pressure increaseand temperature decrease create stress redistribution in the reser-voir-rock matrix, which could induce shear failure. Although tem-
perature decrease has the tendency to create tensile failure, thecombination of pore-pressure and temperature change wouldinduce shear failure. In this paper, we used the Hoek-Brown fail-ure envelope as the criterion for shear failure.
When shear failure occurs, it induces a microseismic event fora fast-slip failure (Zoback et al. 2012). On the other hand, a slow-slip failure could take place but will not create a detectable micro-seismic event. The shear failure is likely to create a large perme-ability enhancement. Zoback et al. (2012) reported that the mainparameters controlling fast- and slow-slip failures are rock me-chanical properties and fracture orientation. High clay and/or or-ganic-material content tends to create a ductile behavior, which islikely to deform rather than break. Thus, shear failure in a high-clay-content (more than 30%) formation is likely to produce slowslip, leading to no detectable microseismic signal.
The Hoek-Brown failure criterion is given as (Parry 2004)
r01N ¼ r03N þffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffimr03N þ Q
q; ð13Þ
where r01N and r03N are normalized maximum and minimum stressdefined as r01=rc and r
03=rc, respectively; rc is unconfined com-
pressive strength; m is Hoek-Brown fitting parameter; Q is rockquality ranging from 0 for jointed rock masses to 1.0 for intactrock; and rc is equal to r1 when r3 is zero.
By rearranging Eq. 13, we can define a new parameter to char-acterize rock failure:
HB ¼ r01N � r03N �ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffimr03N þ 1
q; ð14Þ
where HB is the rock-failure indicator for the Hoek-Brown crite-rion. A positive HB indicates rock failure.
Hoek-Brown failure criterion can be plotted on the Mohr dia-gram with equations described by Parry (2004):
sN ¼ ðcot/0i � cos/0iÞm
8; ð15Þ
where sN is the normalized shear stress defined as s=rc and /0i is
the instantaneous friction angle given as
. . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
150 ft.
400 ft.
650 ft.
Study area
Hydraulic fractures
Dep
th, f
t
0100
200300 400
500600
0100
0
5000
10000
X
Z
Y
Alo
ng h
oriz
onta
l wel
l, ft
Alo
ng h
oriz
onta
l wel
l, ft
Along hydraulic fracture, ft0 100 200 300 400 500
(a) (b)
(c) (d)
600
Along hydraulic fracture, ft
Along
horiz
ontal
well,
ft
Along hydraulic fracture, ft
0 100 200 300 400 500 600
0
50
100
150
200Hydraulic fracture
Fracture network
0
100
200Hydraulic
Fracture network
Overburden
Underburden
Middle Bakken
Fig. 6—A sector model for water-injection study in Bakken: (a) schematic of two multistage hydraulically fractured horizontalwells; (b) schematic of the numerical-model configuration for the reservoir sector and surrounding rocks; and (c) the location ofthe hydraulic fractures, and (d) a natural fracture network in the middle Bakken.
–5000 0 5000 1500.0 2500.0 3000.00
10000
20000
30000
Normal stress, psi
She
ar s
tres
s, p
si
Middle Bakken data
Hoek-Brown failure criterion
m=12.57
Fig. 7—Hoek-Brown failure criterion best fit for a middle Bakkentriaxial test.
902 October 2013 SPE Journal
-
/0i ¼ tan�1 4hcos2 30� þ1
3sin�1h�3=2
� �� 1
�1=2
h ¼ 1þ 16ðmr0nN þ 1Þ
3m2
:
We used a published triaxial test for the middle Bakken forma-tion (Wang and Zeng 2011) to calibrate the model (Fig. 7). Thematched parameter, m, is 12.57. This model was used to analyzethe simulation results in the next section.
Simulation Results
Fig. 8 shows the simulated pressure and water-saturation distribu-tions during water injection. One month after injection, pressure
increases mainly in the fracture network. Later, pressure propa-gates into tight shale matrix. However, the saturation map indi-cates that injected water flows overwhelmingly through thefracture network because the shale matrix is so tight and only mini-mal water can penetrate into the matrix block. In addition, theinjected water reaches the producer before 1 month after injection.This is consistent with observations in the Crescent Energy Point(Wells 2011) field pilot experiments.
Fig. 9 depicts the temperature profiles and the vertical-stresschange. Temperature reduction occurs at the vicinity to the frac-ture network near the injector because energy flux from the reser-voir causes injected water to heat up, thus there is less coolingeffect when it flows deep into the reservoir. After 1 year of injec-tion, the cooling effect takes place inside the tight shale matrixalthough the injected water cannot penetrate into the matrix, as
Time Pressure, psi Water saturation, fraction
1 month
1 year
Alo
n h
oriz
onta
l w
ell,
ft
Along hydraulic fracture, ft0 100 200 300 400 500 600
0
100
200
Hydraulic fracture
Hydraulic fracture
p: 3800 4100 4400 4700 5000
Along hydraulic fracture, ft0 100 200 300 400 500 600
Hydraulic fracture
Hydraulic fracture
sw: 0.3 0.45 0.6 0.75 0.9
Alo
n h
oriz
onta
l w
ell,
ft
Along hydraulic fracture, ft0 100 200 300 400 500 600
0
100
200
Hydraulic fracture
Hydraulic fracture
Alo
ng h
oriz
onta
l wel
l, ft
Alo
ng h
oriz
onta
l wel
l, ft
Along hydraulic fracture, ft0 100 200 300 400 500 600
0
100
200
0
100
200
Hydraulic fracture
Hydraulic fracture
gg
Fig. 8—Simulation results: pressure and water-saturation profiles 1 month and 1 year after cold-water injection.
Time Temperature, oF Stress change in vertical direction, psi 1 month
1 year
Alo
ng h
oriz
onta
l w
ell,
ft
Along hydraulic fracture, ft0 100 200 300 400 500 600
0
100
200
Hydraulic fracture
Hydraulic fracture
T: 110 130 150 170 190 210 230
Along hydraulic fracture, ft0 100 200 300 400 500 600
0
100
200
Hydraulic fracture
Hydraulic fracture
Szz: -7000 -5000 -3000 -1000
Alo
ng h
oriz
onta
l w
ell,
ft
Along hydraulic fracture, ft0 100 200 300 400 500 600
0
100
200
Hydraulic fracture
Hydraulic fracture
Alo
ng h
oriz
onta
l w
ell,
ftA
long
hor
izon
tal
wel
l, ft
Along hydraulic fracture, ft0 100 200 300 400 500 600
0
100
200
Hydraulic fracture
Hydraulic fracture
Fig. 9—Simulation results: Temperature and stress changes in the vertical direction 1 month and 1 year after cold-water injection.
October 2013 SPE Journal 903
-
evidenced in the water-saturation profiles. This cooling effect isdominated by heat conduction, which gradually transfers energyfrom inside the shale matrix to the fracture network.
The vertical-stress profile indicates that the temperaturechange exerts much more effect on the stress change than doesthe pore-pressure change. This is because pressure-change magni-tude is relatively small as the reservoir is injected and producedsimultaneously. On the other hand, the temperature differencebetween the injected water and the reservoir is significant and cancause extensive stress change.
Fig. 10 shows the rock displacement and stress change in theminimum-horizontal-stress direction perpendicular to the hydrau-lic fractures. The hydraulic-fracture-face displacement causes rock
displacement and mechanical strain. It increases the effective stresscaused by compaction. The stress-change magnitude is relativelysmall compared with that of pressure and temperature because thetypical fracture width of a hydraulic fracture in shale reservoirs isnarrow. In this modeling study, we assume the fracture width of 0.2in., which caused a stress buildup of a few hundred psi. After 1year, the cooling effect creates contraction, thus reducing the com-paction. We can see that the temperature change influences thestress change in the minimum-horizontal-stress direction.
Rock-failure indication calculated from Eq. 6 is plotted in Fig. 11.The positive failure indication illustrates rock-failure potential. Itcan be seen from the plot that the failure potential revolves aroundthe area where the cooling effect takes place. The results confirmthe crucial effect of temperature-induced stress on water injectionprocess in the Bakken.
Improved Hydrocarbon Recovery by WaterInjection in Shale Reservoirs
Water injection significantly reduces effective stress, especiallywith cold-water injection into deep or high temperature forma-tions such as the Bakken, as discussed in the previous section.The reduction could improve hydrocarbon recovery in three re-spects. First, it could maintain or even improve permeability ofexisting natural fractures. Second, it could create new microfrac-tures penetrating from the fracture surface into a tight matrixblock, creating flow paths and help produce hydrocarbons fromthe matrix. Finally, injected water could displace hydrocarbons innewly created microfractures.
The proposed process is shown in Fig. 12. Initially (Fig. 12a),liquid hydrocarbon situates inside both fracture and shale matrix,where fractures provide main flow paths and matrix contains themajority of hydrocarbons. Once water injection starts (Fig. 12b),the injected water flows only through fracture networks. This wasobserved in the tracer-testing results during water injection in theBakken, where Crescent Point Energy found immediate waterbreakthrough during the test.
The injected water helps maintain pore pressure inside naturalfractures and lower temperatures (Fig. 12c) in the matrix sur-rounding the fractures, thus reducing effective stress on the frac-tures. Fig. 13 shows laboratory results (Cuisiat et al. 2002) forpermeability of fractures as a function of stress. The reduction of
Time Rock displacement in the direction Stress change in the direction perpendicular to hydraulic fracture, psi perpendicular to hydraulic fracture, ft 1 month
1 year
Alo
ng h
oriz
onta
l w
ell,
ft
Along hydraulic fracture, ft0 100 200 300 400 500 600
0
100
200
Hydraulic fracture
Hydraulic fracture
v: –0.006 –0.003 0 0.003 0.006
Alo
ng h
oriz
onta
l w
ell,
ft
Along hydraulic fracture, ft0 100 200 300 400 500 600
0
100
200
Hydraulic fracture
Hydraulic fracture
Syy: –8000 –6000 –4000 –2000 0
Alo
ng h
oriz
onta
l w
ell,
ft
Along hydraulic fracture, ft0 100 200 300 400 500 600
0
100
200
Hydraulic fracture
Hydraulic fracture
Alo
ng h
oriz
onta
l w
ell,
ft
Along hydraulic fracture, ft0 100 200 300 400 500 600
0
100
200
Hydraulic fracture
Hydraulic fracture
Fig. 10—Simulation results: Rock displacement and stress changes in the direction perpendicular to the hydraulic fracture 1month and 1 year after cold-water injection.
Time Rock failure indicator 1 month
1 year
Alo
ng h
oriz
onta
l w
ell,
ft
Along hydraulic fracture, ft0 100 200 300 400 500 600
0
100
200
Hydraulic fracture
Hydraulic fracture
HB: –0.01 0 0.01
Alo
ng h
oriz
onta
l w
ell,
ft
Along hydraulic fracture, ft0 100 200 300 400 500 600
0
100
200
Hydraulic fracture
Hydraulic fracture
Fig. 11—Simulation results: Likely rock failure sites 1 monthand 1 year after cold-water injection.
904 October 2013 SPE Journal
-
effective stress yields fracture-permeability enhancement and viceversa. It is worth mentioning that typical production in shale res-ervoirs with natural depletion yields high production-rate declineat early production stages, which could be caused by fracture-per-meability reduction as pore pressure in fractures is depleted.
The cooling effect induces contraction and creates tension atthe vicinity of fractures. Once the stress on the surface of matrixexceeds its strength (Fig. 12d), microfractures are formed from thefracture surface into the matrix block. This could happen monthsafter injection because temperatures gradually change duringwater injection. These microfractures create flow paths for hydro-carbons inside the matrix, thus improving the fracture/matrix inter-face area and increasing hydrocarbon production from the matrix.
Fig. 14 shows the experiment conducted by Groisman and Kaplan(1994), in which they investigated the formation of fractures dur-ing desiccation. During the process, the specimens lost their watercontent and contracted, creating tension and forming fractures.The process is generally applicable to the cooling effect thatoccurs during water injection in shale reservoirs.
Moreover, shale material is prone to failure when exposed totemperature variation. Fig. 15 shows the experiment conducted byFu et al. (2004). They observed the influence of heterogeneity onfracture creation during temperature elevation. The results indi-cated that heterogeneity promotes fracture creation during temper-ature alterations. Different materials have different thermal-expansion properties. Raising or reducing temperature creates
Matrix
Macrofractures network
Microfractures
Water flow through fractures
Cooled formation
Temperature-inducedmicrofractures
(a) (b)
(c) (d)
Fig. 12—Enhanced-hydrocarbon-recovery sites resulting from cold-water injection in shale reservoirs: (a) initial condition, (b)early water-injection period where injected water flows only through the fracture network, (c) shale matrix block cooled, and (d)microfractures caused by temperature-induced stress.
Stress dependent initial permeability - Kimmeridge shale
Per
mea
bilit
y (D
arcy
)
Normal stress (MPa)
Induced mated fracture (T1078)
Induced fracture with offset (T1079)
kko
exp(–C= )σ'
Natural fracture (T914)
1,0E+06
1,0E+05
1,0E+04
1,0E+03
1,0E+02
1,0E+01
1,0E+00
1,0E–010 5 10 15 20 25
.n
Fig. 13—Decay of North Sea Kimmeridge shale permeability under varied effective normal stress (Cuisiat et al. 2002).
October 2013 SPE Journal 905
-
strain contrasts between the two connected materials. This mis-match strain can create internal shear stresses and promote fracturecreation. Shale is typically heterogeneous and comprises differentmaterials, such that it is prone to mismatched strains, creating frac-tures during temperature changes.
Conclusions
We developed a practical numerical model that can be used to studythe effect of pore-pressure change and temperature change duringwater injection on stress redistribution in the reservoir and the sur-rounding rocks. Such stress changes have many implications, someof which are summarized in the following conclusions.
In conventional waterflooding, reservoir cooling and pore-pressure increase synergistically promote rock failure, potentiallyreactivating healed natural macrofractures and/or creating newmacrofractures, especially near an injector. The rock cooling cancreate small microfractures on the surface of the matrix blocks.
In shale oil reservoirs, the numerical experiments indicate thatstress changes during water injection can improve oil recovery byopening some of the old macrofractures and creating new micro-fractures perpendicular to the surface of the matrix blocks to pro-mote shallow water invasion into the rock matrix. Then, oil willbe produced from the matrix by countercurrent flow into the frac-
tures, where water will displace the oil from the fractures towardthe production wells.
The new microfractures provide additional interface areabetween macrofractures and matrix to improve oil drainage withsurfactant and CO2 enhanced-oil-recovery techniques. These posi-tive effects are particularly important farther away from the im-mediate vicinity of the hydraulic fracture, where much of theundrained oil resides.
Nomenclature
cp ¼ constant pressure heat capacitycs ¼ heat capacity of solidct ¼ total compressibility, 1/psicv ¼ constant volume heat capacityc/ ¼ pore compressibility, 1/psiCij ¼ rock mechanical properties defined in strain/strain rela-
tion under drained conditionD ¼ depth, ftf ¼ fractional flow, dimensionless
Kb ¼ drained bulk modulus, psim ¼ Hoek-Brown fitting parameter, dimensionlessp ¼ pore pressure, psiq̂ ¼ sink and source term per unit volume, 1/dQ ¼ rock quality, dimensionlessS ¼ phase saturation, dimensionlesst ¼ time, days
T ¼ temperature, T, �RT0 ¼ initial temperature, �F~u ¼ rock displacement vector defined as~u ¼ ½u; v;w�T
u; v; w ¼ rock-displacement components in the x-, y-, and z-directions, ft
~v ¼ Darcy velocity, ft/da ¼ Biot’s poroelastic constant or effective-stress coeffi-
cient, dimensionlessb ¼ coefficient of linear thermal expansion, 1/�F
bfl;t ¼ coefficient of total thermal expansion of fluid, 1/�Fc ¼ pressure gradient, psi/ft
Dri ¼ normal stress change in i-directionDsij ¼ shear stress change in i, j plane
ei ¼ normal strain on i-directioneij ¼ shear strain on i, j-planeev ¼ volume matrix strain, dimensionlessq ¼ density, lbm/ft3r ¼ stress, psir0 ¼ effective stress, psirc ¼ unconfined compressive strength, psit ¼ Poisson’s ratio, dimensionless/ ¼ porosity, dimensionless/0i ¼ instantaneous friction angle, degrees
(a) (b) (c)
Fig. 14—Pictures of an experiment conducted by Groisman and Kaplan (1994) to investigate fracture creation during desiccation.The pictures show the influence of the bottom friction on the size and pattern of created fractures: (a) glass plate, uncoated; (b)bottom coated with 2 mm of grease; and (c) bottom coated with 6 mm of Vaseline. In our case, created microfractures could be sim-ilar to (a) creating small fractures with tight fracture spacing. This is because surface of matrix is under tension caused by coolingeffect; its interior, however, is still at its initial condition.
Fig. 15—A picture of an experiment conducted by Fu et al.(2004) to investigate the influence of heterogeneity on fracturecreation during temperature elevation. Fractures typically formbetween two different materials because the strain mismatch ofthe two materials can create internal shear stress and promotefracture creation.
906 October 2013 SPE Journal
-
Subscripts
o ¼ oil phasew ¼ water phase
Acknowledgments
This work is supported by the US Department of Energy undercontract No. DE-EE0002762. Special thanks to Energy ModelingGroup and Marathon Center of Excellence for Reservoir Studiesat Colorado School of Mines.
References
Axelsson, G., Thorhallsson, S., and Bjonsson, G. 2006. Stimulation of
Geothermal Wells in Basaltic Rock in Iceland. Proc., ENGINE-
Enhanced Geothermal Innovative Network for Europe Workshop, Zur-
ich, Switzerland.
Bjornsson, G. 2004. Reservoir Condition at 3–6 km Depth in the Hellissheidi
Geothermal Field, SW-Iceland, Estimated by Deep Drilling, Cold-Water
Injection and Seismic Monitoring. Proc., the 29th Stanford Geothermal
Workshop, Stanford University, Stanford, California.
Charoenwongsa, S., Kazemi, H., Miskimins, J., et al. 2010. A Fully-
Coupled Geomechanics and Flow Model for Hydraulic Fracturing and
Reservoir Engineering Applications. Paper SPE 137497 presented at
the Canadian Unconventional Resources & International Petroleum
Conference, Calgary, Alberta, Canada, 19–21 October.
Cuisiat, F., Grande, L., and Høeg, K. 2002. Laboratory Testing of Long
Term Fracture Permeability in Shales. Paper SPE 78215 presented at
the SPE/ISRM Rock Mechanics Conference, Irving, Texas, 20–23 Oc-
tober. http://dx. doi.org/10.2118/78215-MS.
Fu, Y.-F., Wong, Y.-L., Poon, C.-S., et al. 2004. Experimental Study of
Micro/Macro Crack Development and Stress-Strain Relations of Cement-
Based Composite Materials at Elevated Temperatures. Cement Concrete
Res. 34 (5): 789–797. http://dx.doi.org/10.1016/j.cemconres.2003.08.029.Ghassemi, A. 2012. A Review of Some Rock Mechanics Issues in Geo-
thermal Reservoir Development. Geotech. Geol. Eng. 30 (3): 647–664.http://dx.doi.org/10.1007/s10706-012-9508-3.
Ghassemi, A., Tarasovs, S., and Cheng, A. H.-D. 2007. A 3-D Study of
the Effects of Thermomechanical Loads on Fracture Slip in Enhanced
Geothermal Reservoirs. Int. J. Rock Mech. Min. 44 (8): 1132–1148.http://dx.doi.org/10.1016/j.ijrmms.2007.07.016.
Ghassemi, A. and Zhang, Q. 2004. Poro-Thermoelastic Mechanisms in
Wellbore Stability and Reservoir Stimulation. Twenty-Ninth Work-
shop on Geothermal Reservoir Engineering, Stanford University, Stan-
ford, California, 26–28 January.
Groisman, A. and Kaplan, E. 1994. An Experimental Study of Cracking
Induced by Desiccation. Europhys. Lett. 25 (6): 415–420. http://dx.doi.org/10.1209/0295-5075/25/6/004.
Kitao, K., Ariki, K., Hatakeyama, H., et al. 1990. Well Stimulation Using
Cold-Water Injection Experiments in the Sumikawa Geothermal Field,
Akita Prefecture, Japan. Proc., 1990 International Symposium on Geo-
thermal Energy: Geothermal Resources Council 1990 Annual Meeting,
Kailua-Kona, Hawaii, Vol. 14, 1219–1224.
Kurtoglu, B., Cox, S. A., and Kazemi, H. 2011. Evaluation of Long-Term
Performance of Oil Wells in Elm Coulee Field. Paper SPE 149273 pre-
sented at the Canadian Unconventional Resources Conference, Alberta,
Canada, 15–17 November. http://dx.doi.org/10.2118/149273-MS.
Nottenburg, R., Rajeshwar, K., Rosenvold, R., et al. 1978. Measurement
of Thermal Conductivity of Green River Oil Shales by a Thermal
Comparator Technique. Fuel 57 (12): 789–795. http://dx.doi.org/10.1016/0016-2361(78)90141-2.
Parry, R. H. G. 2004. Mohr Circles, Stress Paths and Geotechnics, secondedition. New York, New York: Spon Press.
Siratovich, P. A., Sass, I., Homuth, S., et al. 2011. Thermal Stimulation of
Geothermal Reservoirs and Laboratory Investigation of Thermally-
Induced Fractures. Proc., Geothermal Resources Council Annual
Meeting 2011, San Diego, California, Vol. 35, 1529–1535.
Tarasovs, S. and Ghassemi, A. 2012. On the Role of Thermal Stress in
Reservoir Stimulation. Oral presentation given at the Thirty-Seventh
Workshop on Geothermal Reservoir Engineering, Stanford University,
Stanford, California, 30 January–February 1.
Vernik, L. and Landis, C. 1996. Elastic Anisotropy of Source Rocks:
Implications for Hydrocarbon Generation and Primary Migration.
AAPG Bull. 80 (4): 531–544. http://dx.doi.org/10.1306/64ED8836-
1724-11D7-8645000102C1865D.
Wang, C. and Zeng, Z. 2011. Overview of Geomechanical Properties of
Bakken Formation in Williston Basin, North Dakota. Paper ARMA
11-199 presented at the 45th U.S. Rock Mechanics/Geomechanics
Symposium, San Francisco, Califonia, 26–29 June.
Wells, P. 2011. Bakken Pilot Success Could See Waterflood Tested in
Others Tight Oil Plays, Daily Oil Bulletin, April 19, 2011, http://
pdf.dailyoilbulletin.com/110419/dob-letter-2011-04-19.pdf (accessed 10
December 2011).
Wood, T. and Milne, B. 2011. Waterflood Potential Could Unlock Billions
of Barrels: Crescent Point Energy: Reported by Dundee Securities
Ltd., http://www.investorvillage.com/uploads/44821/files/CPGdundee.
pdf (accessed 4 May 2012).
Ye, F. 2010. Sensitivity of Seismic Reflections to Variations in Anisotropy
in the Bakken Formation, Williston Basin, North Dakota. MS thesis,
University of Texas at Austin, Texas (June 2010).
Zoback, M. D., Kohli, A., Das, I., et al. 2012. The Importance of Slow
Slip on Faults During Hydraulic Fracturing Stimulation of Shale Gas
Reservoirs. Paper SPE 155476 presented at the SPE Americas Uncon-
ventional Resources Conference, Pittsburgh, Pennsylvania, 5–7 June.
http://dx.doi.org/10.2118/155476-MS.
Appendix A—Mechanical Properties
Isotropic material can be understood as
C11 C12 C13
C21 C22 C23
C31 C32 C33
C44
C55
C66
2666666664
3777777775
¼
2Gþ k k kk 2Gþ k kk k 2Gþ k
2G
2G
2G
2666666664
3777777775;
� � � � � � � � � � � � � � � � � � � ðA-1Þ
where G is shear modulus and k is the Lamé coefficient.In orthotropic material, the Young’s modulus, shear modulus,
and Poisson’s ratio are direction dependent. For such material,Eq. 13 takes the form
C11 C12 C13
C21 C22 C23
C31 C32 C33
C44
C55
C66
2666666664
3777777775
¼
1=Ex�tyx�Ey �tzx=Ez
�tyx�Ey 1�Ey
�tzy�Ez�tzx=Ez
�tzy�Ez 1=Ez1�2Gxy
1�2Gyz
1=2Gzx
266666666666664
377777777777775
�1
� � � � � � � � � � � � � � � � � � � ðA-2Þ
October 2013 SPE Journal 907
-
where Ei is Young modulus in the i-direction, tij is Poisson’s ratioin the i, j plane, and Gij is shear modulus in the i, j plane.
Appendix B—Input Data for Numerical-SimulationStudies
See Tables B-1 and B-2.
TABLE B-1—FIVE-SPOT WATERFLOOD MODEL
Parameter Value Unit Parameter Value Unit
Rock properties Rock-mechanics properties
Porosity 0.2 — Elastic modulus 3�106 psiVertical permeability 10 md Poisson’s ratio 0.25 —
Horizontal permeability 50 md Shear modulus 106 psi
Biot’s constant 0.8 —
Thermal properties Linear thermal expansion 10�5 �F�1
Rock heat capacity 0.21 Btu/(lbm-�F)
Rock thermal conductivity 35 Btu/(D-ft-�F) Initial reservoir conditions
Rock density 170 lbm/ft3 Initial pressure 2,700 psi
Initial temperature 200 �F
Well control Initial water saturation 0.25 —
Injected temperature 80 �F
Injection rate 500 B/D 160-acre spacing for full pattern
Quarter of five-spot, injector/producer distance is 1,867 ft
TABLE B-2—A SECTOR MODEL OF WATER INJECTION IN THE BAKKEN
Parameter Value Unit Parameters Value Unit
Well Control Fluid Properties
Injected temperature 80 �F Oil
Total injection rate 1,040 B/D API gravity1 42 �API
Sector injection rate 40 B/D Viscosity1 0.36 —
Maximum injection pressure 5,000 psi Heat capacity 0.6 Btu/(lbm-�F)
Total production rate 1,040 B/D Heat conductivity 8.0 Btu/(D-ft-�F)
Sector production rate 40 B/D Water —
Minimum production pressure 1,000 psi Viscosity 0.5 —
Heat capacity 1.0 Btu/(lbm-�F)
Heat conductivity 9.0 Btu/(D-ft-�F)
Middle Bakken Rock Properties
Flow Properties Rock-Mechanics Properties
Porosity1 0.08 — Elastic modulus3 4.5�106 psiEffective permeability1 0.05 md Poisson’s ratio3 0.30 —
Matrix permeability 0.001 md Shear modulus5 1.7�106 psiFracture permeability 150 md Biot’s constant3 0.8 —
Linear thermal expansion 10�5 �F�1
Thermal Properties
Heat capacity 0.21 Btu/(lbm��F) Initial ConditionsThermal conductivity2 16.8 Btu/(D-ft-�F) Initial pressure3 4,500 psi
Density 170 lbm/ft3 Initial temperature1 240 �F
Initial water saturation 0.25 —
Initial Stress Conditions
Vertical stress3 8,000 psi
Maximum horizontal stress3 6,800 psi
Minimum horizontal stress3 6,000 psi
Underburden and Overburden Rock Properties
Flow Properties Rock-Mechanics Properties in Horizontal Direction
Porosity 0.01 — Drained elastic modulus4 4.5�106 psiTotal permeability 0 md Poisson’s ratio4 0.30 —
Shear modulus5 1.7�106 psiThermal Properties Biot’s constant 0.8 —
Heat capacity 0.21 Btu/(lbm-�F) Linear thermal expansion 10�5 �F�1
908 October 2013 SPE Journal
-
Perapon Fakcharoenphol is a PhD candidate in petroleum en-gineering at the Colorado School of Mines (CSM). His researchinterests are numerical simulation of coupled geomechanicsand transport in porous and fractured media. He worked as areservoir engineer for PTT Exploration and Production Com-pany (PTTEP) from 2002 to 2008. Fakcharoenphol holds an MSdegree in petroleum engineering from Imperial College, Lon-don, and a BS degree in petroleum engineering from Chula-longkorn University, Thailand.
Sarinya Charoenwongsa is a petroleum engineer at ChevronEnergy Technology Company. She worked as a reservoir engi-neer for PTTEP from 2002 to 2007 and as a post-doctoralresearch scientist at CSM in 2012. She holds BS and MS degreesin chemical engineering from King’s Mongkut Institute of Tech-nology, Ladkrabang, and Chulalongkorn University, respec-tively. Charoenwongsa holds a PhD degree in petroleumengineering from CSM.
Hossein Kazemi is the Chesebro’ Distinguished Professor ofPetroleum Engineering at CSM and codirector of the Mara-thon Center of Excellence for Reservoir Studies (MCERS).Kazemi is a member of the National Academy of Engineer-ing and is an SPE Distinguished Member and Honorary Mem-
ber. He retired from Marathon Oil Company in 2001 afterserving as the Director of Production Research, the Managerof Reservoir Technology, and an executive technical fellow.At CSM, Kazemi teaches graduate courses and supervisesresearch in reservoir modeling, well testing, and improvedoil- and gas-recovery processes. He holds BS and PhDdegrees in petroleum engineering from the University ofTexas at Austin.
Yu-Shu Wu is a professor and Foundation CMG ResearchChair in Petroleum Engineering at the CSM. He is also a guestscientist at the earth sciences division of the Lawrence Berke-ley National Laboratory (LBNL). At CSM, he is teaching petro-leum reservoir engineering courses, supervising graduatestudents, and conducting research in the areas of multi-phase fluid and heat flow in porous media, enhanced-oil-re-covery operation, CO2 geosequestration, reservoirsimulation, enhanced geothermal systems, and unconven-tional hydrocarbon reservoirs. Wu was a staff scientist withthe LBNL, from 1995 to 2008. He started his career as a petro-leum engineer at RIPED, Beijing. Wu holds BS (Eqv.) and MSdegrees in petroleum engineering in China and MS and PhDdegrees in reservoir engineering from the University of Califor-nia at Berkeley.
TABLE B-2 (continued)—A SECTOR MODEL OF WATER INJECTION IN THE BAKKEN
Parameter Value Unit Parameters Value Unit
Thermal conductivity2 16.8 Btu/(D-ft-�F)
Density 170 lbm/ft3 Rock-Mechanics Properties in Vertical Direction
Elastic modulus3 6.8�106 psiPoisson’s ratio3 0.18 —
Shear modulus 1.7�106 psiBiot’s constant 0.8 —
Linear thermal expansion 10�5 �F�1
1 Kurtoglu et al. (2011)2 Nottenburg et al. (1978)3 Wang and Zeng (2011)4 Ye (2010)5 Calculated from other mechanical properties.
October 2013 SPE Journal 909